Do attention and memory explain the performance of software developers?

Writing and modifying source code are core activities in software development and evolution. The outcome of a coding task in terms of quality may depend on several aspects, such as the difficulty of the task or the complexity of the system. Besides, it is well known that individual characteristics of developers, like the programming experience, play a lead role in this. Recent work started exploring the influence that cognitive human aspects have on the ability of developers to acquire information from the source code ( e.g. , finding security blind spots). However, it is still unknown to what extent such aspects influence their ability of completing coding tasks. In this paper, we theorize that two cognitive human aspects, attention and memory, play a role in predicting the outcome of a coding task. We conducted a controlled experiment involving 32 participants (18 bachelor students, 9 master students, 2 Ph.D. students. and 3 practitioners), in which we asked them to complete two bug-fixing and two feature implementation tasks. We measured, for each of them, three attention-related factors ( i.e., alerting, orienting, and executive control) and two memory-related ones ( i.e., working memory and immediate recall) through well-established psychometric tests. Finally, we investigated to what extent these factors can explain the correctness, the readability and the time taken to complete a task in function of such factors. Our results show that all the attention- and memory-related factors achieved very low correlation with correctness and time. Indeed, the number of years of programming experience is far more important than all the other variables we considered for explaining the correctness and the time required to complete a task. Moreover, we found a significant relationship between orienting (an attention-related factor) and code readability.